Identification of the chaotic behavior of Lorenz equations using vector support machines

Introduction: The article is derived from the research Characterization of complex signals with Computational Intelligence techniques that has been ongoing since 2016 within the investigation Group Modelacion en Ingenieria de Sistemas MIS of the Universidad Distrital Francisco Jose de Caldas. Objective: Determine the chaotic behavior in equations of Lorenz by using data directly taken from the time domain without prior processing, in order to classify a chaotic system. Methodology: Firstly, through simulation training data is acquired; later, it is used validation data to observe the system response. Finally, it is presented a discussion together with a set of conclusions regarding the data obtained. Results: In most of the implemented vector support machines a positive classification is prevailing. Conclusion: The data set used for the classification of the chaotic behavior in Lorenz equations was achieved implementing vector support machines, so they may be an alternative to obtaining behavior classification where data are directly taken from the time domain with none prior processing. Originality: This paper is expected to serve to further developments as in diagnosis of patients using biological signals. This work is aimed to the observation of the characteristics manifested in the vector support machines to chaotic system classification. Limitations: None preliminary data processing is performed wherefore such classification is subjected by the values obtained directly from the simulation.